ATG9B

ATG9B is a tissue-specific functional ortholog of yeast Atg9p that drives autophagosome biogenesis by delivering membrane lipids to the growing phagophore (PMID:15755735, PMID:37938170). It assembles as a conserved homotrimer that functions as a lipid scramblase, recapitulates the subcellular trafficking and steady-state distribution of ATG9A, can compensate for ATG9A loss in starvation-induced autophagy, and forms a heteromeric complex with the lipid-transfer protein ATG2A (PMID:37938170). In autophagosome formation it acts upstream of cargo docking, facilitating recruitment of LC3 and p62-associated ubiquitinated proteins to nascent phagophores, such that its loss sensitizes cells to ER stress-induced death through accumulation of ubiquitinated substrates (PMID:28740555). Beyond canonical autophagy, ATG9B has distinct context-dependent roles: it is stabilized by and trafficked with MYH9 to the cell edge, where it accelerates focal adhesion assembly by promoting integrin β1–Talin-1 interaction and integrin β1 activation to drive cancer invasion in an autophagy-independent manner (PMID:34131310); it controls actin depolymerization during bacterial internalization under ULK1 kinase control, with its loss causing accumulation of actin filaments and phosphorylated LIM kinase and cofilin (PMID:38706859); and it localizes to mitochondria via an N-terminal targeting sequence, where its expression collapses mitochondrial membrane potential, promotes mtDNA release, and triggers apoptosis (PMID:41811769). ATG9B expression is set by multiple transcriptional and post-transcriptional inputs, including suppression by several exosomal miRNAs and activation through ASCL2 and Wnt/β-catenin signaling in cancer stemness and drug resistance (PMID:28740555, PMID:35971776, PMID:35882624, PMID:41495834). A homozygous frameshift truncating the C-terminal cytosolic domain causes a rare human neurodevelopmental disorder, and the truncated protein is unstable and fails to reach peripheral vesicles, defining the C-terminal domain as required for peripheral trafficking.

1. 2005 Medium

   Established that ATG9B is a functional ortholog of yeast Atg9p, answering whether this paralog participates in the autophagy machinery at all.

   Evidence Fluorescent-tagged localization, ATG9A siRNA knockdown rescue, and LC3 colocalization in HeLa cells

   PMID:15755735

   Open questions at the time

   • Molecular activity of the protein not defined
   • No structural information
   • Direct binding partners unidentified
2. 2017 Medium

   Placed ATG9B upstream of cargo capture by showing its phagophores dock LC3 and p62-associated ubiquitinated proteins, explaining why its loss causes ER stress-induced death.

   Evidence Hepatocarcinogenesis mouse model with siRNA knockdown and LC3/p62/ubiquitin immunofluorescence

   PMID:28740555

   Open questions at the time

   • Mechanism of p62/ubiquitin recruitment not defined at molecular level
   • Single lab
3. 2021 High

   Revealed an autophagy-independent function: ATG9B drives focal adhesion assembly and cancer invasion through MYH9-stabilized trafficking and integrin β1 activation.

   Evidence Reciprocal Co-IP, domain mapping (aa368-411), STUB1 ubiquitination assays, and focal adhesion/integrin β1–Talin-1 assays

   PMID:34131310

   Open questions at the time

   • Whether this function requires the scramblase activity unknown
   • Generality beyond colorectal cancer untested
4. 2023 High

   Defined the biochemical activity and quaternary structure of ATG9B, establishing it as a homotrimeric lipid scramblase that binds ATG2A and compensates for ATG9A.

   Evidence Single-particle cryo-EM, in vitro lipid scramblase assay, ATG9A knockout complementation, and ATG2A co-immunoprecipitation

   PMID:37938170

   Open questions at the time

   • Stoichiometry/architecture of the ATG9B-ATG2A complex not resolved structurally
   • Tissue context of scramblase function untested in vivo
5. 2024 Medium

   Connected ATG9B to cytoskeletal control during infection, showing it promotes actin depolymerization under ULK1 regulation to govern bacterial internalization.

   Evidence ATG knockout screen, siRNA knockdown, bacterial internalization assays, actin/pLIMK/pCofilin immunofluorescence, ULK1 inhibition in HeLa cells

   PMID:38706859

   Open questions at the time

   • Direct effectors linking ATG9B to LIMK/cofilin unknown
   • Whether ULK1 phosphorylates ATG9B directly untested
6. 2024 Medium

   Linked ATG9B to human disease and mapped a trafficking determinant, showing a C-terminal frameshift causes neurodevelopmental disease and abolishes peripheral vesicle targeting.

   Evidence Patient genetics, knock-in mouse model, and cell-based localization/stability comparison of truncated vs WT protein (preprint)

   Open questions at the time

   • Preprint, single lab
   • Mechanistic basis of neurodevelopmental phenotype not established
   • How the C-terminal domain mediates peripheral trafficking unknown
7. 2026 Medium

   Uncovered a pro-apoptotic mitochondrial role distinct from ATG9A, mediated by an N-terminal mitochondrial targeting sequence.

   Evidence Tagged ATG9B expression, mitochondrial membrane potential and mtDNA release assays, apoptosis readouts, and N-terminal domain truncation in tumor cell lines

   PMID:41811769

   Open questions at the time

   • Whether endogenous ATG9B localizes to mitochondria physiologically unclear
   • Mechanism of membrane potential collapse undefined
   • Single lab

• How ATG9B's single scramblase activity is partitioned across its diverse autophagic, focal-adhesion, actin-remodeling, and mitochondrial-apoptotic functions, and how its trafficking is switched between these fates, remains unresolved.
• No unified model linking scramblase activity to autophagy-independent roles
• Regulators selecting perinuclear vs peripheral vs mitochondrial localization unknown